Merge commit '1c7f40bf8322' from llvm.org/main into next

Author: git apple-llvm automerger

llvm/lib/Target/X86/X86ISelLowering.cpp

@@ -45172,6 +45172,18 @@ bool X86TargetLowering::isGuaranteedNotToBeUndefOrPoisonForTargetNode(
   case X86ISD::Wrapper:
   case X86ISD::WrapperRIP:
     return true;
+  case X86ISD::PACKSS:
+  case X86ISD::PACKUS: {
+    APInt DemandedLHS, DemandedRHS;
+    getPackDemandedElts(Op.getSimpleValueType(), DemandedElts, DemandedLHS,
+                        DemandedRHS);
+    return (!DemandedLHS ||
+            DAG.isGuaranteedNotToBeUndefOrPoison(Op.getOperand(0), DemandedLHS,
+                                                 PoisonOnly, Depth + 1)) &&
+           (!DemandedRHS ||
+            DAG.isGuaranteedNotToBeUndefOrPoison(Op.getOperand(1), DemandedRHS,
+                                                 PoisonOnly, Depth + 1));
+  }
   case X86ISD::INSERTPS:
   case X86ISD::BLENDI:
   case X86ISD::PSHUFB:
@@ -45242,6 +45254,10 @@ bool X86TargetLowering::canCreateUndefOrPoisonForTargetNode(
   case X86ISD::BLENDI:
   case X86ISD::BLENDV:
     return false;
+  // SSE packs.
+  case X86ISD::PACKSS:
+  case X86ISD::PACKUS:
+    return false;
   // SSE target shuffles.
   case X86ISD::INSERTPS:
   case X86ISD::PSHUFB:

llvm/test/CodeGen/X86/combine-pack.ll

@@ -26,22 +26,19 @@ define <8 x i16> @combine_packss_v4i32_signsplat(<4 x i32> %a0, <4 x i32> %a1) {
   ret <8 x i16> %signsplat
 }
 
-; TODO: Failure to remove unnecessary signsplat through freeze
 define <8 x i16> @combine_packss_v4i32_freeze_signsplat(<4 x i32> %a0, <4 x i32> %a1) {
 ; SSE-LABEL: combine_packss_v4i32_freeze_signsplat:
 ; SSE:       # %bb.0:
 ; SSE-NEXT:    pcmpgtd %xmm1, %xmm0
 ; SSE-NEXT:    pcmpeqd %xmm1, %xmm1
 ; SSE-NEXT:    packssdw %xmm1, %xmm0
-; SSE-NEXT:    psraw $15, %xmm0
 ; SSE-NEXT:    retq
 ;
 ; AVX-LABEL: combine_packss_v4i32_freeze_signsplat:
 ; AVX:       # %bb.0:
 ; AVX-NEXT:    vpcmpgtd %xmm1, %xmm0, %xmm0
 ; AVX-NEXT:    vpcmpeqd %xmm1, %xmm1, %xmm1
 ; AVX-NEXT:    vpackssdw %xmm1, %xmm0, %xmm0
-; AVX-NEXT:    vpsraw $15, %xmm0, %xmm0
 ; AVX-NEXT:    retq
   %cmp = icmp sgt <4 x i32> %a0, %a1
   %ext = sext <4 x i1> %cmp to <4 x i32>